Reference-based cell type matching of spatial transcriptomics data

Abstract

With the advent of multiplex fluorescence in situ hybridization (FISH) and in situ RNA sequencing technologies, spatial transcriptomics analysis is advancing rapidly. Spatial transcriptomics provides spatial location and pattern information about cells in tissue sections at single cell resolution. Cell type classification of spatially-resolved cells can also be inferred by matching the spatial transcriptomics data to reference single cell RNA-sequencing (scRNA-seq) data with cell types determined by their gene expression profiles. However, robust cell type matching of the spatial cells is challenging due to the intrinsic differences in resolution between the spatial and scRNA-seq data. In this study, we systematically evaluated six computational algorithms for cell type matching across four spatial transcriptomics experimental protocols (MERFISH, smFISH, BaristaSeq, and ExSeq) conducted on the same mouse primary visual cortex (VISp) brain region. We find that while matching results of individual algorithms vary to some degree, they also show agreement to some extent. We present two ensembl meta-analysis strategies to combine the individual matching results and share the consensus matching results in the Cytosplore Viewer (https://viewer.cytosplore.org) for interactive visualization and data exploration. The consensus matching can also guide spot-based spatial data analysis using SSAM, allowing segmentation-free cell type assignment.

Optimizing Xenium In Situ data utility by quality assessment and best practice analysis workflows

“…We therefore set out to benchmark Xenium against available datasets from the same general region of the mouse brain, primarily isocortex. The imaging-based SRT datasets included the recently published high sensitivity in situ sequencing (HS-ISS) 24 , STARmap PLUS 23 , sequential smFISH and MERFISH from the Allen Institute 25 , ExSeq 22 , BaristaSeq 26 , osmFISH 18 and the Vizgen MERFISH Mouse Brain Receptor Map. For the sequencing-based SRT, we had three different publicly available Visium datasets.…”

SpaceWalker: Interactive Gradient Exploration for Spatial Transcriptomics Data

“…To quantitatively evaluate to what extent the top-ranked genes detected by the spatial and HD filters corresponded to known layerspecific marker genes in the mouse visual cortex 25 , we calculated a frequency count for filtered genes by counting how often genes end up in the top two ranks at different seed cell locations per cortical layer.…”

“…To this end, we used two publicly available ST datasets of the mouse visual cortex that were acquired in the context of the SpaceTx consortium 24 . We selected these datasets (smFISH and MERFISH) because of the abundance of literature on the laminar spatial architecture of the mouse visual cortex 25 . The SpaceTx datasets contain manual annotation of the cortical layers, which enables investigating the robustness of SpaceWalker against ST protocols applied to the same tissue sample (smFISH and MERFISH).…”

“…Cells located near the boundary of the spatial layer are less likely to be in the same layer as the manual seed cell annotation, and thus show a lower LP value. Another explanation for the lower LP values could be due to the fact that transcriptomically similar cells may appear in multiple layers 25 . This cell type cross-talk across cortical layer boundaries explains why flood-fill neighborhood searching seeded close to the manually annotated layer boundaries may span across the manually annotated layer boundaries.…”

“…All layer-specific marker genes have a high correlation score with their corresponding layer annotation, proving that the correlation metric identifies genes with contrasting patterns within a specific layer. Known marker genes as reported 25 are presented in the top of the heatmaps, and the other genes are grouped based on layer and correlation score. Genes with a high frequency count in a specific layer also exhibit a high correlation score with the manually annotated layer masks, demonstrating the ability of the filters to capture local spatial variability.…”